---
base_model: roneneldan/TinyStories-33M
library_name: Distily
tags:
- generated_from_trainer
model-index:
- name: distily_bench_obj_cross_v2.7
  results: []
---

# distily_bench_obj_cross_v2.7

This student model is distilled from the teacher model [roneneldan/TinyStories-33M](https://huggingface.co/roneneldan/TinyStories-33M) using the dataset (unspecified).

The [Distily](https://github.com/lapp0/distily) library was used for this distillation.

It achieves the following results on the evaluation set:
- eval_enwikippl: 167.9577
- eval_frwikippl: 36527.4492
- eval_zhwikippl: 179398.4219
- eval_tinystoriesppl: 10.7634
- eval_loss: 1.3170
- eval_runtime: 6.482
- eval_samples_per_second: 77.137
- eval_steps_per_second: 9.719

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment.

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed
-->

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- distillation_objective: DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl, layer_mapper=None, projector=None), hs_loss_component=LossComponent(label=hs, weight=0, loss_fn=None, layer_mapper=None, projector=None), attn_loss_component=LossComponent(label=attn, weight=0, loss_fn=None, layer_mapper=None, projector=None))
- train_embeddings: True
- learning_rate: 0.004
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1.0

### Resource Usage
Peak GPU Memory: 8.0557 GB

### Eval-Phase Metrics
| step | epoch | enwikippl | frwikippl | loss | runtime | samples_per_second | steps_per_second | tinystoriesppl | zhwikippl |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| **teacher eval** |  | 169.9865 | 47377.9414 |  |  |  |  | 3.9789 | 4998.1294 |
| 0 | 0 | 21812.5137 | 63073.9414 | 6.0880 | 6.5014 | 76.907 | 9.69 | 11544.7471 | 48756.5938 |
| 500 | 0.0808 | 207.5609 | 68082.6719 | 1.4006 | 6.4908 | 77.032 | 9.706 | 11.2631 | 330400.9375 |
| 1000 | 0.1616 | 175.8123 | 42073.3594 | 1.3250 | 6.5141 | 76.756 | 9.671 | 10.6938 | 234691.875 |
| 1500 | 0.2424 | 169.7431 | 37693.0312 | 1.3185 | 6.4776 | 77.19 | 9.726 | 10.7252 | 183317.4531 |
| 2000 | 0.3232 | 168.7467 | 36424.6914 | 1.3180 | 6.4817 | 77.14 | 9.72 | 10.8563 | 177257.4688 |
| 2500 | 0.4040 | 168.3811 | 36910.1914 | 1.3173 | 6.4798 | 77.163 | 9.723 | 10.8075 | 184790.6562 |
| 3000 | 0.4848 | 166.7585 | 36342.6992 | 1.3171 | 6.4837 | 77.116 | 9.717 | 10.6836 | 179207.0781 |
| 3500 | 0.5656 | 168.1269 | 36651.1523 | 1.3173 | 6.4924 | 77.014 | 9.704 | 10.7683 | 179589.9688 |
| 4000 | 0.6464 | 168.4202 | 36404.2031 | 1.3173 | 6.4897 | 77.045 | 9.708 | 10.7906 | 180599.0 |
| 4500 | 0.7272 | 167.8602 | 36342.6992 | 1.3173 | 6.5499 | 76.338 | 9.619 | 10.7915 | 178348.4531 |
| 5000 | 0.8080 | 167.6392 | 36548.0430 | 1.3173 | 6.5041 | 76.875 | 9.686 | 10.7287 | 180261.9219 |
| 5500 | 0.8888 | 168.0618 | 36527.4492 | 1.3174 | 6.5137 | 76.761 | 9.672 | 10.7701 | 179302.6406 |
| 6000 | 0.9696 | 168.0098 | 36527.4492 | 1.3170 | 6.4907 | 77.034 | 9.706 | 10.7678 | 179398.4219 |
| 6188 | 1.0 | 167.9577 | 36527.4492 | 1.3170 | 6.482 | 77.137 | 9.719 | 10.7634 | 179398.4219 |

### Framework versions
- Distily 0.2.0
- Transformers 4.44.0
- Pytorch 2.3.0
- Datasets 2.21.0